Search Results (12,063)

Accurate detection of oil and gas (O&G) facilities in high-resolution remote sensing imagery is critical for infrastructure surveillance and sustainable resource management, yet conventional detectors struggle with severe class imbalance, extreme scale variation, and arbitrary orientation. In this work, we propose OGF Oriented R-CNN (Oil and Gas Facility Detection Oriented Region-based Convolutional Neural Network), an enhanced oriented detection model derived from Oriented R-CNN that integrates three improvements: (1) O&G Loss Function, (2) Class-Aware Hard Example Mining (CAHEM) module, and (3) Feature Pyramid Network with Feature Enhancement Attention (FPNFEA). Working in synergy, they resolve the coupled challenges more effectively than any standalone fix and do so without relying on rigid one-to-one matching between modules and individual issues. Evaluated on the O&G facility dataset comprising 3039 high-resolution images annotated with rotated bounding boxes across three classes (well sites: 3006, industrial and mining lands: 692, drilling: 244), OGF Oriented R-CNN achieves a mean average precision (mAP) of 82.9%, outperforming seven state-of-the-art (SOTA) models by margins of up to 27.6 percentage points (pp) and delivering a cumulative gain of +10.5 pp over Oriented R-CNN. Full article

Electric vehicles (EVs) rely on a battery pack as their primary energy source, making it a critical component for their operation. To guarantee safe and correct functioning, a Battery Management System (BMS) is employed, which uses variables such as State of Charge (SOC) to set charge/discharge limits and to monitor pack health. In this article, we propose a Multilayer Perceptron (MLP) network to estimate the SOC of a 14.8 V battery pack installed in a robotic vacuum cleaner. Both offline and online (real-time) tests were conducted under continuous load and with rest intervals. The MLP’s output is compared against two commonly used approaches: NARX (Nonlinear Autoregressive Exogenous) and CNN (Convolutional Neural Network). Performance is evaluated via statistical metrics, Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE), and we also assess computational cost using Operational Intensity. Finally, we map these results onto a Roofline Model to predict how the MLP would perform on an automotive-grade microcontroller unit (MCU). A generalization analysis is performed using Transfer Learning and optimization using MLP–Kalman. The best performers are the MLP–Kalman network, which achieved an RMSE of approximately 13% relative to the true SOC, and NARX, which achieved approximately 12%. The computational cost of both is very close, making it particularly suitable for use in BMS. Full article

Traditional fusion methods for integrating multi-source gravity data rely on predefined mathematical models that inadequately capture complex nonlinear relationships, particularly at wavelengths shorter than 10 km. We developed a convolutional neural network incorporating differential marine geodetic data (DMGD-CNN) to enhance marine gravity anomaly recovery from HY-2A satellite altimetry. The DMGD-CNN framework encodes spatial gradient information by computing differences between target points and their surrounding neighborhoods, enabling the model to explicitly capture local gravity field variations. This approach transforms absolute parameter values into spatial gradient representations, functioning as a spatial high-pass filter that enhances local gradient information critical for short-wavelength gravity signal recovery while reducing the influence of long-wavelength components. Through systematic ablation studies with eight parameter configurations, we demonstrate that incorporating first- and second-order seabed topography derivatives significantly enhances model performance, reducing the root mean square error (RMSE) from 2.26 mGal to 0.93 mGal, with further reduction to 0.85 mGal achieved by the differential learning strategy. Comprehensive benchmarking against international gravity models (SIO V32.1, DTU17, and SDUST2022) demonstrates that DMGD-CNN achieves 2–10% accuracy improvement over direct CNN predictions in complex topographic regions. Power spectral density analysis reveals enhanced predictive capabilities at wavelengths below 10 km for the direct CNN approach, with DMGD-CNN achieving further precision enhancement at wavelengths below 5 km. Cross-validation with independent shipborne surveys confirms the method’s robustness, showing 47–63% RMSE reduction in shallow water regions (<2000 m depth) compared to HY-2A altimeter-derived results. These findings demonstrate that deep learning with differential marine geodetic features substantially improves marine gravity field modeling accuracy, particularly for capturing fine-scale gravitational features in challenging environments. Full article

Farmland shelterbelts are crucial elements in safeguarding agricultural ecological security and sustainable development, with their precise extraction being vital for regional ecological monitoring and precision agriculture management. However, constrained by their narrow linear distribution, complex farmland backgrounds, and spectral confusion issues, traditional remote sensing methods encounter significant challenges in terms of accuracy and generalization capability. In this study, six representative deep learning semantic segmentation models—U-Net, Attention U-Net (AttU_Net), ResU-Net, U²-Net, SwinUNet, and TransUNet—were systematically evaluated for farmland shelterbelt extraction using high-resolution Gaofen-6 imagery. Model performance was assessed through four-fold cross-validation and independent test set validation. The results indicate that convolutional neural network (CNN)-based models show overall better performance than Transformer-based architectures; on the independent test set, the best-performing CNN model (U-Net) achieved a Dice Similarity Coefficient (DSC) of 91.45%, while the lowest DSC (88.86%) was obtained by the Transformer-based TransUNet model. Among the evaluated models, U-Net demonstrated a favorable balance between accuracy, stability, and computational efficiency. The trained U-Net was applied to large-scale farmland shelterbelt mapping in the study area (Alar City, Xinjiang), achieving a belt-level visual accuracy of 95.58% based on 385 manually interpreted samples. Qualitative demonstrations in Aksu City and Shaya County illustrated model transferability. This study provides empirical guidance for model selection in high-resolution agricultural remote sensing and offers a feasible technical solution for large-scale and precise farmland shelterbelt extraction. Full article

Pollen identification is critical for melissopalynology (honey authentication), ecological monitoring, and allergen tracking, yet manual microscopic analysis remains labor-intensive, subjective, and error-prone when multiple grains overlap in realistic samples. Existing automated approaches often fail to address multi-grain scenarios or lack systematic comparison across classical and deep learning paradigms, limiting their practical deployment. This study proposes a subinstance-based classification framework combining YOLOv12n object detection for grain isolation, independent classification via classical machine learning (ML), convolutional neural networks (CNNs), or Vision Transformers (ViTs), and majority voting aggregation. Five classical classifiers with systematic feature selection, three CNN architectures (ResNet50, EfficientNet-B0, ConvNeXt-Tiny), and three ViT variants (ViT-B/16, ViT-B/32, ViT-L/16) are evaluated on four datasets (full images vs. isolated grains; raw vs. CLAHE-preprocessed) for four berry pollen species (Ribes nigrum, Ribes uva-crispa, Lonicera caerulea, and Amelanchier alnifolia). Stratified image-level splits ensure no data leakage, and explainable AI techniques (SHAP, Grad-CAM++, and gradient saliency) validate biological interpretability across all paradigms. Results demonstrate that grain isolation substantially improves classical ML performance (F1 from 0.83 to 0.91 on full images to 0.96–0.99 on isolated grains, +8–13 percentage points), while deep learning excels on both levels (CNNs: F1 = 1.000 on full images with CLAHE; ViTs: F1 = 0.99). At the instance level, all paradigms converge to near-perfect discrimination (F1 ≥ 0.96), indicating sufficient capture of morphological information. Majority voting aggregation provides +3–5% gains for classical methods but only +0.3–4.8% for deep models already near saturation. Explainable AI analysis confirms that models rely on biologically meaningful cues: blue channel moments and texture features for classical ML (SHAP), grain boundaries and exine ornamentation for CNNs (Grad-CAM++), and distributed attention across grain structures for ViTs (gradient saliency). Qualitative validation on 211 mixed-pollen images confirms robust generalization to realistic multi-species samples. The proposed framework (YOLOv12n + SVC/ResNet50 + majority voting) is practical for deployment in honey authentication, ecological surveys, and fine-grained biological image analysis. Full article

Recent advances in deep learning have had a significant impact on biomedical applications, driving precise actions in automated diagnostic processes. However, integrating neural networks into medical devices requires meeting strict requirements regarding computing power, energy efficiency, reconfigurability, and latency, essential conditions for real-time inference. Field-Programmable Gate Array (FPGA) architectures provide a high level of flexibility, performance, and parallel execution, thus making them a suitable option for the real-world implementation of machine learning (ML) and deep learning (DL) models in systems dedicated to the analysis of physiological signals. This paper presents a review of intelligent algorithms for electrocardiogram (ECG) signal classification, including Support Vector Machines (SVMs), Artificial Neural Networks (ANNs), Recurrent Neural Networks (RNNs), Long Short-Term Memory Networks (LSTMs), and Convolutional Neural Networks (CNNs), which have been implemented on FPGA platforms. A comparative evaluation of the performances of these hardware-accelerated solutions is provided, focusing on their classification accuracy. At the same time, the FPGA families used are analyzed, along with the reported performances in terms of operating frequency, power consumption, and latency, as well as the optimization strategies applied in the design of deep learning hardware accelerators. The conclusions emphasize the popularity and efficiency of CNN architectures in the context of ECG signal classification. The study aims to offer a current overview and to support specialists in the field of FPGA design and biomedical engineering in the development of accelerators dedicated to physiological signals analysis. Full article

Background/Objectives: The early and accurate identification of impacted teeth in the maxilla is critical for effective dental treatment planning. Traditional diagnostic methods relying on manual interpretation of radiographic images are often time-consuming and subject to variability. Methods: This study presents a deep learning-based approach for automated classification of impacted maxillary canines using panoramic radiographs. A comparative evaluation of four pre-trained convolutional neural network (CNN) architectures—ResNet50, Xception, InceptionV3, and VGG16—was conducted through transfer learning techniques. In this retrospective single-center study, the dataset comprised 694 annotated panoramic radiographs sourced from the archives of a university dental hospital, with a mildly imbalanced representation of impacted and non-impacted cases. Models were assessed using accuracy, precision, recall, specificity, and F1-score. Results: Among the tested architectures, VGG16 demonstrated superior performance, achieving an accuracy of 99.28% and an F1-score of 99.43%. Additionally, a prototype diagnostic interface was developed to demonstrate the potential for clinical application. Conclusions: The findings underscore the potential of deep learning models, particularly VGG16, in enhancing diagnostic workflows; however, further validation on diverse, multi-center datasets is required to confirm clinical generalizability. Full article

Achieving comprehensive improvements in the drying rate (DR) and the quality after drying of agricultural products is a major goal in the field of drying. To further shorten the drying time while improving product quality, this study introduced a Convolutional Neural Network (CNN) and MultiHead Attention (MHA) to enhance the prediction accuracy of the Long Short-Term Memory (LSTM) network regarding the properties of dried samples. These properties included DR, shrinkage rate (SR), and total color difference (ΔE). The CNN-LSTM-MHA network was proposed, developing a novel hot-air drying (HAD) scenario utilizing an intelligent temperature control system based on the real dynamics of material properties. The results of drying experiments with temperature-sensitive yuba showed that the CNN-LSTM-MHA network’s predictive accuracy was better than that of other networks, as evidenced by its coefficient of determination (R²: 0.9855–0.9999), root mean square error (RMSE: 0.0001–0.0099), and mean absolute error (MAE: 0.0001–0.0120). Comparative analysis with fixed-temperature drying indicated that CNN-LSTM-MHA-controlled drying significantly reduced drying time and enhanced the SR, color, rehydration ratio (RR), texture, protein content, fat content, and microstructure of yuba. Overall, the findings highlight the potential of CNN-LSTM-MHA-based intelligent drying as a viable strategy for yuba stick processing, providing insights for other food drying applications. Full article

The crude oil futures market is highly susceptible to policy changes and international relations, which often trigger abrupt jumps in prices. The existing literature rarely considers jump volatility and the underlying impact mechanisms. This study proposes a hybrid forecasting model integrating a convolutional neural network (CNN) and self-attention (Transformer) for high-frequency financial data, based on the complex network characteristics between trading information and multi-market financialization indicators. Empirical results demonstrate that incorporating complex network indicators enhances model performance, with the CNN–Transformer model with a complex network achieving the highest predictive accuracy. Furthermore, we verify the model’s effectiveness and robustness in the WTI crude oil market via Diebold–Mariano tests and external event shock. Notably, this study also extends the analytical framework to jump intensity, thereby providing a more accurate and robust jump forecasting model for risk management and trading strategies in the crude oil futures market. Full article

The existing multimodal image classification methods often suffer from several key limitations: difficulty in effectively balancing local detail and global topological relationships in hyperspectral image (HSI) feature extraction; insufficient multi-scale characterization of terrain features from light detection and ranging (LiDAR) elevation data; and neglect of deep inter-modal interactions in traditional fusion methods, often accompanied by high computational complexity. To address these issues, this paper proposes a comprehensive deep learning framework combining convolutional neural network (CNN), a graph convolutional network (GCN), and wavelet transform for the joint classification of HSI and LiDAR data, including several novel components: a Spectral Graph Mixer Block (SGMB), where a CNN branch captures fine-grained spectral–spatial features by multi-scale convolutions, while a parallel GCN branch models long-range contextual features through an enhanced gated graph network. This dual-path design enables simultaneous extraction of local detail and global topological features from HSI data; a Spatial Coordinate Block (SCB) to enhance spatial awareness and improve the perception of object contours and distribution patterns; a Multi-Scale Elevation Feature Extraction Block (MSFE) for capturing terrain representations across varying scales; and a Bidirectional Frequency Attention Encoder (BiFAE) to enable efficient and deep interaction between multimodal features. These modules are intricately designed to work in concert, forming a cohesive end-to-end framework, which not only achieves a more effective balance between local details and global contexts but also enables deep yet computationally efficient interaction across features, significantly strengthening the discriminability and robustness of the learned representation. To evaluate the proposed method, we conducted experiments on three multimodal remote sensing datasets: Houston2013, Augsburg, and Trento. Quantitative results demonstrate that our framework outperforms state-of-the-art methods, achieving OA values of 98.93%, 88.05%, and 99.59% on the respective datasets. Full article

The growing demand for digital content protection has significantly increased the importance of image watermarking, particularly in light of the rising vulnerability of multimedia content to unauthorized modifications. In recent years, research has increasingly focused on leveraging deep learning architectures to enhance watermarking performance, addressing challenges related to transparency, robustness, and payload capacity. Numerous deep learning-based watermarking methods have demonstrated superior effectiveness compared to traditional approaches, particularly those based on Convolutional Neural Networks (CNNs), Generative Adversarial Networks (GANs), Transformers, and diffusion models. This paper presents a comprehensive survey of recent developments in both conventional and deep learning-based image watermarking techniques. While traditional methods remain prevalent, deep learning approaches offer notable improvements in embedding and extraction efficiency, particularly when facing complex attacks, including those generated by advanced AI models. Applications in areas such as deepfake detection, cybersecurity, and Internet of Things (IoT) systems highlight the practical significance of these advancements. Despite substantial progress, challenges remain in achieving an optimal balance between invisibility, robustness, and capacity, particularly in high-resolution and real-time scenarios. This study concludes by outlining future research directions toward develop robust, scalable, and efficient deep learning-based watermarking systems capable of addressing emerging threats in digital media environments. Full article

Human Activity Recognition (HAR) using wearable sensors enables impactful applications in healthcare, fitness, and smart environments, but it also faces challenges related to data privacy, non-independent and identically distributed (non-IID) data, and limited computational resources on edge devices. This study proposes an efficient and personalized federated learning (PFL) framework for HAR that integrates federated training with model compression and per-client fine-tuning to address these challenges and support deployment on resource-constrained devices (RCDs). A convolutional neural network (CNN) is trained across multiple clients using FedAvg, followed by magnitude-based pruning and float16 quantization to reduce model size. While personalization and compression have previously been studied independently, their combined application for HAR remains underexplored in federated settings. Experimental results show that the global FedAvg model experiences performance degradation under non-IID conditions, which is further amplified after pruning, whereas per-client personalization substantially improves performance by adapting the model to individual user patterns. To ensure realistic evaluation, experiments are conducted using both random and temporal data splits, with the latter mitigating temporal leakage in time-series data. Personalization consistently improves performance under both settings, while quantization reduces the model footprint by approximately 50%, enabling deployment on wearable and IoT devices. Statistical analysis using paired significance tests confirms the robustness of the observed performance gains. Overall, this work demonstrates that combining lightweight model compression with personalization providing an effective and practical solution for federated HAR, balancing accuracy, efficiency, and deployment feasibility in real-world scenarios. Full article

Individual animal identification is a cornerstone of animal welfare practices and is of crucial importance for food safety and the protection of humans from zoonotic diseases. It is also a key prerequisite for enabling automated processes in modern dairy farming. With newly emerging technologies, visual animal identification based on machine learning offers a more efficient and non-invasive method with high automation potential, accuracy, and practical applicability. However, a common challenge is the limited variability of training datasets, as images are typically captured in controlled environments with uniform backgrounds and fixed poses. This study investigates the impact of foreground segmentation and background removal on the performance of convolutional neural networks (CNNs) for cow identification. A dataset was created in which training images of dairy cows exhibited low variability in pose and background for each individual, whereas the test dataset introduced significant variation in both pose and environment. Both a fine-tuned CNN backbone and a model trained from scratch were evaluated using images with and without background information. The results demonstrate that although training on segmented foregrounds extracts intrinsic biometric features, background cues carry more information for individual recognition. Full article

Background/Objectives: Conventional workflows, peripheral blood smears, and bone marrow assessment supplemented by LDI-PCR, molecular cytogenetics, and array-CGH, are expert-driven in the face of biological and imaging variability. Methods: We propose an AI pipeline that integrates convolutional neural networks (CNNs) and transfer learning-based models with two explainable AI (XAI) approaches, LIME and Grad-Cam, to deliver both high diagnostic accuracy and transparent rationale. Seven public sources were curated into a unified benchmark (66,550 images) covering ALL, AML, CLL, CML, and healthy controls; images were standardized, ROI-cropped, and split with stratification (80/10/10). We fine-tuned multiple backbones (DenseNet-121, MobileNetV2, VGG16, InceptionV3, ResNet50, Xception, and a custom CNN) and evaluated the accuracy and F1-score, benchmarking against the recent literature. Results: On the five-class task (ALL/AML/CLL/CML/Healthy), MobileNetV2 achieved 97.9% accuracy/F1, with DenseNet-121 reaching 97.66% F1. On ALL subtypes (Benign, Early, Pre, Pro) and across tasks, DenseNet121 and MobileNetV2 were the most reliable, achieving state-of-the-art accuracy with the strongest, nucleus-centric explanations. Conclusions: XAI analyses (LIME, Grad-CAM) consistently localized leukemic nuclei and other cell-intrinsic morphology, aligning saliency with clinical cues and model performance. Compared with baselines, our approach matched or exceeded accuracy while providing stronger, corroborated interpretability on a substantially larger and more diverse dataset. Full article

Weed competition in red beet (Beta vulgaris L. Conditiva Group) directly reduces crop yield and quality, making detection and eradication essential. This study proposed a three-phase experimental protocol for multi-class detection (cultivation and six types of weeds) based on RGB (red-green-blue) colour images acquired in a greenhouse, using state-of-the-art deep learning (DL) models (YOLO and RT-DETR family). The objective was to evaluate and optimise performance by identifying the combination of architecture, model scale and input resolution that minimises false negatives (FN) without compromising robust overall performance. The experimental design was conceived as an iterative improvement process, in which each phase refines models, configurations, and selection criteria based on performance from the previous phase. In phase 1, the base models YOLOv9s and RT-DETR-l were compared at 640 × 640 px; in phase 2, the YOLOv8s, YOLOv9s, YOLOv10s, YOLO11s, YOLO12s and RT-DETR-l models were compared at 640 × 640 px and the best ones were selected using the F1 score and the FN rate. In phase 3, the YOLOv9 (s = small, m = medium, c = compact, e = extended) and YOLOv10 (s = small, m = medium, l = large, x = extra-large) families were scaled according to the number of parameters (s/m/c-e/l-x sizes) and resolutions of 1024 × 1024 and 2048 × 2048 px. The best results were achieved with YOLOv9e-2048 (F1: 0.738; mAP@0.5 (mean Average Precision): 0.779; FN: 28.3%) and YOLOv10m-2048 (F1: 0.744; mAP@0.5: 0.775; FN: 27.5%). In conclusion, the three-phase protocol allows for the objective selection of the combination of architecture, scale, and resolution for weed detection in greenhouses. Increasing the resolution and scale of the model consistently reduced FNs, raising the sensitivity of the system without affecting overall performance; this is agronomically relevant because each FN represents an untreated weed. Full article